Professional Documents
Culture Documents
Gitedment depth compared with: equation (5.4) (after Bode & Roik (1987)) HEADED ANCHORS IN-UNCRACKED CONCRETE, ‘Standardization Institute (TGL). Some variations, such as the use of different factors (AUS, 1985), may indicate a different interpretation of the predictive capacity of equation (9.3). Use of a pyramidal shape for the failure surface for single anchors (TGL) and anchor groups (PCI and TGL) simplifies the determination of A, in cases involving complex geometries. '5,2.3.2. Method according to Bode & Hanenkamp (1985) and Bode & Roike (1987). Bode & Hanenkamp (1985) and Bode & Roik (1987) present an empirical equation for the average ultimate tension load for single-headed anchors exhibiting concrete-cone failures based on the results of over 100 tests with headed studs. The anchorages tested had an embedment depth range of approximately 40—140mm. Concrete cube strengths f., of 20-45 MPa Were measured at the time of testing. Also recorded at the. time of testing were the corresponding cylinder and splitting tensile strengths fz and fe- Regression analysis of the test data gave the expressions for mean strength 0.96 hes [1 F (dn/hes)] (fee)? (N= (S.4a) Nomieginde = U-B9A? [1 + (dnl) (6) — NY 6.40) Figure 5.11 shows the normalized pull-out strengths N, [fa a8 a function of the anchorage length yyq. Curves for the average pull-out strength Nan according to equation (5.4) are plotted against the data using two boundary values for the ratio of head diameter to embedment depth dy ee “The results of the regression analysis were employed to derive an expression for the characteristic strength (5% fractile, 90% probability) Nexto) = B.DARE [L + (dlls) ee)? (N) (5.5a) Nuxcoytosey = 9-65 AY? TL + Gn/her) 1 (6)°F ®) (6.56) Comparison of equations (5.4) and (5.5) shows the characteristic strength to be ~0.81 times the mean value. Factors to account for the effects of spacing and edge distance are based Nace) roMONOTONIC LOADING IN CONCRETE * os cs ==: 1 - A | 8 oO i 8 7 | ~ | 8 S 5 Equation 6.8) z oe (7 equation (58) 5 os 3 2 2’ ; V8 = [se ti acoring equation (4) Y : zi 7 BB ee) 10 | | as 08 08 70 os 88 70 eh he Fig. 5.12. Test data for headed anchor groups exhibiting concrete-cone failures compared vith. ‘equation (5.6), plotted as a function of actual-te-critical ‘spacing ratio (ajter Bode & Roik (1987) 46 on an approach originally developed by Bligehausen & Pusill-Wachtsmuth (1982) (Fig. 5.12). Two boundary conditions are assumed for anchor spacing. (@) For a theoretical centre-to-centre anchor spacing of s = 0, the capacity of'a multiple anchor fastening is reduced to the capacity of one anchor. (®) If the anchor spacing s is greater than or equal to a value (4f,)) taken as critical for development of the full concrete-cone failure load, then the group attains its maximum tension capacity, which is n times the failure load of a single anchor. ‘A straight line is assumed between these limiting values, resulting in Mamll + sa— DIG he) 6.6) where n is the number of studs acting together as one group, s is the centre~ to-centre stud spacing and N,,, is the mean failure load of a single anchor according to equation (5.4). Figure 5.12 shows the results of tests with headed stud groups (n = 2, 3, 4 and 9 anchors) loaded in tension through a stiff loading plate (Bode & Roik, 1987). The measured failure loads have been normalized to n times the failure load of a single anchor by use of equation (5.4), and are ploited as a function of the ratio of actual spacing to the assumed critical value of 4hy. The straight line functions for predicted capacity as given by equation (6.6) are plotted against the data in Fig. 9.12. ‘Similarly, a critical edge distance c,, necessary for the development of the full concrete-cone failure load is considered. Two values for the critical edge distance are assumed, depending on the number of proximate free edges. These assumptions give Nasnetge = MerNam Num (6.7) where cy = LShy (one free edge), Cy = 20h (two or more free edges, e.g. anchor in a comer) and Nj.» is the mean failure load of a single anchor (Grom equation (5.4)). Figure 5.13 plots the results of tests conducted with headed studs located near a free edge (Bode & Roik, 1987). The measured failure loads have been normalized using values calculated from equation (5.4), and are plotted in Ne mgrouyFi 5.13. Test data for headed anchors located near an edge compared with equation (5.7), plotted as a fimetion of the ratio of edge distance to embedment depth (after Bode & Roik f1987)) we HEADED ANCHORS IN UNCRACKED CONCRETE relation to ratio of edge distance to embedment depth c/hy. The bilinear relationship given by equation (5.7) is shown for comparison. 5.2.3.3. Method according to Eligehausen, Fuchs & Mayer (1987/1988) and Rehm et al. (1988) (J method). The ¥ method for predicting the strength of headed anchors exhibiting concrete-cone failures was first proposed by Lehmann (1984). It was revised by Pusill-Wachismuth (1982), Eligehausen & Pusill-Wachtsmuth (1982), and Riemann (1985), and developed to the form presented here by Eligehausen et al, (1987/1988) and Rehm et al. (1988). The method is generally referred to in the source literature as the x method. It has been renamed the y method to conform with CEB notation. | Eligehausen et al. (1987/1988) and Rehm et al. (1988) present @ ‘comprehensive procedure for the calculation of ultimate tension loads for headed anchors exhibiting concrete-cone failures, based on the results of 196 tests with headed studs, The tests involved anchorage depths of 40-525 mm and measured concrete strength f., of 20-60 MPa, Based on a regression ‘analysis ofthe test results, the foliowing relations were obtained for the mean strength Numicate) 1s.5mye fe (NY (5.8a) Nambia = 1TH AEE (5.80) Figure 5.14 shows a histogram of the ratio of measured to predicted failure Joads (Rehim et ai., 1988). The coefficient of variation is given as 14% and the 5% fractile of the failure load is derived as 0.77 times the average value. ‘The equations for characteristic strength (5% fractile, 90% probability) are thus given as Nexon = 12hH fe? ON) 6.98) Nexioinsey = 13hG? F25 ON) (6.96) ‘As with the method described in ACI 349, equation (5.8) assumes the failure Joad to depend solely on the concrete tensile strength (taken as proportional to f9%) and the embedment depth. i in Fig, 5.15 (Rehm et al., 1988), test results for headed studs exhibiting concrete-cone failures are plotted as a function of embedment depth hi. The ‘curve for mean ultimate strength given by equation (5 8a) is plotted against the data for comparison. The tests were carried out in concretes of varying strengths; the measured failure loads have been normalized to a concrete strength fi, = 25MPa using the square root of the compression strength 08; NtesiNam os Equation (5.7) Namacoording to equation (5.4) 7s het ”1200 ~ 4 ke 225A na = NA2sit 1000} - 2001 — —f~$— ‘according 10 equation (5.8) for headed anchors exhibiting concrete-cone failures: test results from Fig. 5.14. Ratio of actual to predicted tension capacity ag Single anchors without edge influences (afier Rehm ob et al., 1988) Fig. 5.15 {above right). Comparison of equation (6.8) with tension test results for headed anchors exhibiting concrete-cone failures normalized t0 a ‘concrete strength f.. = 25MPa and plotied as a function of embedment depth (after Rehm et ala, 1988) 48 Nyt = 18h! See°9/1000 | 600 |— 400) - aoe tA 200 a0 00 ratio. As such, Fig. 5.15 does not indicate the accuracy of the assumed relationship between failure load and concrete strength (N, & Vf.) inherent in equation (5.8), but it does indicate the dependency of N, on hy? as derived from the regression analysis. ‘AS with the method of Bode & Hanenkamp (1985), allowance for the effects of edge distance and spacing is based on the approach originally developed by Eligehausen & Pusill-Wachtsmuth (1982). Based on an assumed crack angle for the failure cone of « ~ 55°, and supposing that overlap of adjacent failure cones leads to a reduction in group capacity, the investigators proposed the critical spacing s,, = 3h (Cf. sa, = 4h, (Bode & Roik, 1983)) to avoid any spacing influence. A straight line relationship is assumed between the boundary conditions of s., = 3 and s., = 0, which results in the following ‘equation for the mean strength of a ‘group’ of two anchors Nasmgrup = Vso (5.10) wo Dt Shey 52 6.1 where Jay = Shy and N,.. is the mean failure load of a single anchor from equation (5.8). Figure 5.16 (Rehm ef al., 1988) plots test results for headed anchor groups (two anchors) loaded in tension through a stiff loading plate as a function of anchor spacing divided by embedment depth s/hye. All tests ended in cconcrete-cone failure. The ultimate loads obtained from the tests have been divided by the calculated failure load for a single anchor by use of equation (5.8) to normalize the results with respect to concrete strength and to facilitate comparison with the bilinear relationship given by equation (5.10). For the calculation of the ultimate load of orthogonally arranged quadruple fastenings (see Fig. 5.17), Rehm et al. (1988) proposed that factors yj, be separately derived for each orthogonal direction and multiplied as follows Nemaiop = Wai ¥aNom 6.12)| Fig. 5.16. Comparison of equation (5.10) with test results for headed stud groups (two anchors) loaded in tension through a stiff loading plate: test results have been divided by the calculated faiture load for a Single anchor using ‘equation (5.8) and plovted (a5 a function of anchor spacing divided by ‘mbedment depth (after Rehm et al., 1988) © Fig, 5.17, Concrete-cone failure of a quadruple ‘fastening HEADED ANCHORS IN,UNCRACKED CONCRETE Va = 1+ 5h5q) $2 (8.13) where 5; is the centre-to-centre spacing measured in direction i and Nm is the mean failure load of a single anchor from equation (5.8). Figure 5.18 (Rehm et al., 1988) plots the tension failure loads from tests on quadruple fastenings with equal spacing in two directions as a function of anchor spacing divided by embedment depth. Tests with headed, expansion and undercut anchors are included. As with Fig. 5.16, the data have been normalized with the failure load for a single fastening calculated from equation (6.8). Equation (5.12) data are plotted against the test data for comparison. Implicit in equation (5.12) is the assumption that negative effects due to spacing (ie. overlapping of stress fields) and other factors (e.g. edge distance) on load capacity can be accounted for by multiplication of individual effects determined through measurement or analysis.’ This assumption of a multiplicative effect forms the basis for the ¥ method, Rehm ef al. (1988) extend this idea to orthogonally and symmetrically arranged multiple fastenings with a maximum spacing in each direction s,, < sq by substituting the distance between the outermost anchors of the group 5, for s; in equation (6.13) and maltiplying all terms as follows None = Th VesNe 6.14) where Nag is the average failure load of a multiple anchor group with anchors, Nm is the mean failure load of a single anchor from equation (5.8) 8 T T Num according to equation (6.8) el ° | jt 3 & 3 Equation (6.1 2 :quation (6.10) 1 oy | ait ©: Anchor 1 Headed etud 1 2 3 thotPenn eee eee MONOTONIC LOADING IN CONCRETE Fig. 5.18. Comparison of ‘equation (5.12) with test results for quadruple anchor 3 ‘groups loaded in tension 4 - through a stiff loading 7 plate: test results have been ‘Nugn aocording to onustion (5.8) e divided by the calculated ° Jailure load for a single anchor using equation (5.8) ‘and plotted as a function of ‘anchor spacing divided by embedment depth (after Rehm et al., 1988) C : N&: ’ 3 a) [ese] eunton 6.12) Lo Anchor . 1B Headed stud 4 sitet predicted failure load for groups of headed anchors &s.a fsction of the spacing ofthe outermost anchors ° § (afer Eligehausen et al, e, 19%) Fig. 5.19. Ratio of actual 10 E bee. aroun Meal grour o Fauds 2 see 2 testucs © 36 studs fiat 185 mm ‘Mxcakgroup aocording to equation (5. sane S28 Fig. 5.20. Model for the influence of load ‘eccentricity on the failure load for a group of two infinitely stiff anchors loaded through a rigid plate (according to Riemann, 1985) Nya WeDo Nye = Woven My2= Nam Went + alse 52 Vo= tt + 2etse) 50HEADED ANCHORS IN UNCRACKED CONCRETE Bog = AF SiulSee S Mi 6.15) where s,, is the distance between outermost anchors in direction i (see Fig. 5.19), n-is the number of anchors in direction i, m = 1 for anchors in one tow and m, = 2 for anchors in two directions. As before, all anchors in the group are assumed to be loaded equally. Figure 5.19 (Eligchausen et al., 1992) plots the ratio of actual to predicted tension capacity of groups of headed anchors in relation to the spacing of the outermost anchors s,. Group fastenings with up to 36 anchors were loaded concentrically in tension through a rigid loading plate to assure equal oad distribution. The embedment depth A was held constant at 185 mm. ‘The spacing of the outer anchors was 100-875 mm; the spacing of the individual anchors was 100-400mm (0.54h.-—2.2he). The concrete strength f, was ~25MPa. All anchorages exhibited concrete-cone failures. Figure 5.20 shows the influence of load eccentricity on the carrying behaviour of anchor groups. With centric loading (case 1), equation (5.14) is valid. When the load acts on one headed stud (case IN), the limit load of the group corresponds to the value for an individual anchor, irrespective of the axial distance. For a tensile force acting anywhere within the group (case I), a hyperbolic curve between the limiting cases I and IIT can be assumed for the failure load of the anchor group. The influence of eccentricity ‘on the rupture load of the group can thus be determined by use of an additional coefficient Y. Numa = ¥e¥eNem 6.16) where N,,., is the mean failure load of a single anchor from equation (5.8) ve = WEL + Gels) = 1 6.17) where e is the eccentricity of tension force from the centroid of anchors (see Fig. 5.20); ¢ < 5/2. Equation (5.16) can also be used for groups of anyprevent’splitting. faire of an anchorage in a reinforced concrete member. ‘While the work of Hasselwander et al. (1987) included tests with anchors in specially. reinforced specimens, the testing was limited and no attempt was made to isolate the influence of the reinforcement on the anchorage: strength. Although an equation was presented for the strength of the attchoragés'tested, it must be assumed that this equation is valid only for the reinforcing arrangements tested. Rehini et al: (1988) have proposed that splitting failures can be avoided in anreinforced members with concrete strength f. greater than 20 MPa.if edge distance > 1.5hy, spacing = 3.0hy and member thickness = 2.0 hey. 5.3.1. Types of failure and load—displacement behaviour ‘The failure’ modes of headed anchors loaded in shear can be categorized as follows (see Fig. 5.37) 6MONOTONIC LOADING IN CONCRETE Fig. 5.37, Failure modes of headed anchors loaded in shear Fig. 5.38. Shelt-shaped spall and associated shaft deformation of a headed stud subjected to shear loading (from Fuchs, 1990) a 5.3.1.1. Steel failure (Fig. 5.37(a)). For deeper embedments, failure is generally characterized by bending of the anchor shaft. Due to the locally high pressure in front of the bolt, a conchiform concrete spall may occur near the upper surface of the concrete before maximum load is obtained. With increasing load, the anchor shaft yields and/or fractures. Depending ‘on the concrete strength and the size, embedment and steel qualities of the anchor, deformations at failure may be relatively large. The failure load increases with increasing steel strength and cross-sectional area. 5.3.1.2. Concrete failure (Fig. 5.37(6)— (g)). Failure is characterized by one of three possible modes. (@) Concrete spalting is shown in Fig, 5.37(a) (see also Fig. 5.38). Various authors define the onset of concrete spalling as a point of failure for the anchorage based on serviceability considerations. Particularly for deeper anchorages, the maximum load may be reached at much greater displacements and. anchor deformations. (©) Lateral.concrete-cone failures due to edge proximity are shown in Fig. 5.370). ‘The original of the lateral cone may be at the head of the1 i j Fig. 5.39. Typical load— displacement curves of headed anchors under tension and shear loading (after Hawkins (1987)) HEADED ANCHORS IN: UNCRACKED CONCRETE Loa kN + shear (ky): ste0t fare —O— Shear {kt concrete fale Displacement. mm anchor for short stiff embedments; for deeper embedments it usually begins closer to the surface of the concrete. If the development of the full concrete cone is limited by spacing or edge condition (Fig. 5.37(c)—@)) ot by the depth of the member (Fig. 5.37(f), the failure load is generally reduced. As with tension loading, the failure load is influenced by the concrete tensile capacity. Other influencing factors include the side cover c, the flexural stiffness of the anchor shaft and the embedment depth. (© Concrete crushing combined with fracture of the concrete behind the anchor and Subsequent pull-out of the anchor is shown in Fig. 5.37(g). ‘This failure mode generally occurs with simall embedment depths (fer 4d—6d) or anchor groups with small spacing. 5.3.1.3. The typical load—displacement bchavionr of a single headed stud loaded in ‘pure’ shear can be described as follows (see Fig. 5.39). On initial loading, the shaft of the anchor beats directly on the surface concrete, this being the stiffest load path. As the surface concrete is crushed, flexural stresses are generated in the anchor shaft as it attempts to redistribute the load over aa greater depth. Increasing load will usually lead to the formation of a shell- shaped spall (see Fig. 5.38), the depth of which is determined primarily by the flexural stiffness of the anchor shaft. Spalling increases the flexural demand on the anchor shaft and simultaneously reduces the internal level arm, For shallow anchors, the onset of spalling may be accompanied by a flexural/shear failure of the anchor shaft or concrete fracture originating at the anchor head similar to that shown in Fig. 5.37(g). In the case of headed anchors with sufficiently large embedment depth and shaft strength, however, the load may continue to increase until the anchor is sheared off through a combination of shear, flexural and tensile stresses. ‘The displacements associated with this post-spalling behaviour may be quite large, and in general the mode of shear load transfer (concrete crushing and spalling followed by anchor shaft bending) leads to larger displacements at failure than those associated with tension loading. In the case of a headed anchor loaded in the direction of a free edge and exhibiting « lateral concrete-cone failure, the ultimate load (and corresponding displacement) increases with increasing edge distance. For a given anchor, however, the load associated with the failure mode described in section 5.3.1.1 (palling followed by steel failure) represents an upper bound on the anchorage shear capacity 6 |MONOTONIC LOADING IN CONCRETE Table 5.1. Shear equations and values of the shear reduction factor 5.3.2. Failure loads — steel failure Headed anchors subjected to shear loads experience shear, bending and axial stresses. There is no generally accepted theoretical formulation for the calculation of the rupture load. It is often assumed that failure occurs when the bending stresses exceed the steel tensile strength: the bending moment has been calculated for the case of a beam on an elastic or elasto-plastic support Gasler & Witta, 1967; Cziesielski & Friedmann, 1983; Friberg, 1940; Wiedenroth, 1971). The rupture loads thus determined, however, rarely match the test results (Fuchs, 1984). It seems, therefore, that empirical formulations are more useful at present. ‘The literature presents several methods that consider the failure load solely 1s a function of the steel strength, and one method that includes the concept of shear—friction. These approaches are reviewed in sections 5.3.2.1 and 53.2.2. ‘5.3.2.1. Method according to ACI 349-85 (1985) (shear—friction theory). ‘The shear capacity for embedments is calculated from Ve= AS, (bf, N) 6.31) where ¢ is a strength reduction factor (@ = 0.55 for structural shapes, fabricated steel sections and shear lugs embedded in the concrete, and = (0.85 for bolts, studs and bars), A, is the total cross-sectional area of headed anchors attached to embedments, f, is the steel yield strength (= 120 kgffin’ (~828MPa)) and y. is the coefficient of friction. The term ‘embedment’ as used by ACI 349-85 refers to an anchor plate with anchors either welded or bolted to the plate. The coefficient of friction is typically 4 = 0.9 for concrete or grout against as-rolled steel with the contact plane a full baseplate thickness below the concrete surface, « = 0.7 for concrete or group against as-rolled steel with the contact plane coincidental with the concrete surface, and = 0.55 for grouted conditions with the contact plane between grout and as-rolled steel exterior to the concrete surface. Tf the friction coefficient is taken as 0.7 for steel against concrete, the ultimate shear strength ( = 1.0) of the embedment is given by Veo = 0.7 Arf, (bf, N) 6.32) ‘5.3.2.2; Methods not based'on shear—friction theory. Many investigators have found that the average steel failure load of a headed anchor loaded in shear can be predicted by Visa Af, (bf, N) (5.33) ‘Various forms of this equation in the literature relate the sheat reduction factor ato either the ultimate or the yield strength of the steel. Table 5.1 gives Reference ‘Test basis | ‘Shear equation « Burdette et al. (1987) Steelleonerete Van = adh 0.65 ‘Chesson, Faustino & Munse (1965) Steel/steel OAS, 0.55—-0.64 ‘Fuchs (1984) ‘Steel/concrete i OAS, 06 Hencky, Huber & Mises Theory Vin = adh, 0.58 ADT (1985) Steel/concrete Vom = OAs, 0.7 Kiingnet, Mendonca & Malik (1982) Steel/conerete Vins adift 0.675 Kulak, Fisher & Struik (1987) ‘Steel/stee! Vin = OAS, 0.62 McMackin, Slutter & Fisher (1973) Steel/conerete Vin = oh 10 Roik (1982) Steel/concrete Ven = oA, 07 ‘Shoup er al, (1963) ‘Stee/conerete Vom = OAS, 1.0 TVA (1975) ‘Steel/grout/concrete Vin = Af, 0.53 Valtinat (1982) Screws Jom = OAS, 0.625 6Fig. 5.40, Comparison of ‘actual failure loads for anchors loaded in shear towards a free edge and values predicted from equation (5.34) (after Klingner & Mendonca, 19826) HEADED ANCHORS IN-UNCRACKED CONCRETE these equations with the cortesponding shear reduction factors proposed by the authors. "The calculated strength of a group of headed anchors exhibiting steel failure depends on the distribution of load to the anchors. In general, when equal load distribution is assured (as when studs are welded to a concentrically- loaded anchor plate), the failure load is » times the value obtained from equation (5.28) (n = number of fastening elements). Fuchs (1990) investigated the case of two anchors projecting through an anchor plate loaded in shear with variable clearances between the holes in the anchor plate, and the anchor shafts. Expansion anchors and undercut anchors manufactured from high-strength brittle steels were employed in the tests, which showed reductions in group strength of up (0 20% for hole clearance differences of up to 4mm. The degree of reduction is related to the ability of the anchors to redistribute load through plastic deformation of the anchor shafts. 7 Cook & Klingner (1989) propose that the shear strength in a multiple-anchor connection should be taken as 50% of the tensile strength (V,/T, = 0.50). ‘The shear strength (steel failure) of undercut anchors in a multiple-anchor connection should be taken as 60% of the tensile strength (V,/T, = 0.60). 5.3.3. Failure loads — lateral concrete-cone failure (edge failure) In the case of headed anchors loaded towards a free edge, a lateral cone failure may pre-empt failure of the steel (see Fig. 5.40). According to Fuchs & Eligehausen (1986a) and Stichting Bouwresearch (1971), the angle of fracture as measured from a line passing through the anchor and parallel to the edge is approximately 30°—35? on average, and the depth of the lateral failure cone is approximately 1.3—1.5 times the edge distance. Klingner & Mendonca (1982b) report the angle as ~ 45° for edge distances of 125mm or more, decreasing linearly to 25° as the edge distance approaches zero. ACI 349-85 (1985) assumes an angle of 45° ‘The ultimate load capacity associated with this failure mode depends on the behaviour of concrete under multiaxial stress. As in the case of tension loading, a complete analytical determination of the concrete failure load is not yet available. The following equations arc therefore derived empirically, taking account of observed behaviour. As in the case of tension loading, these fracture angle measurements are the basis for the spacing requirements associated With the various methods described in sections 5.3.3.1-5.3.3.4. ‘The methods are compared in sections 5.3.3.5—5.3.3.7. Other factors are discussed in sections 5.3.3.8 and 5.3.3.9. 5.3.3.1. Method according to ACI 349-85 (1985). The average strength of an anchor exhibiting @ lateral concrete-cone failure is calculated by assuming a uniform stress of 4(/2)"* Ibffin’ on the projected area of the failure cone Vom = G40) Ae (bf 6.34) (Hom = 90.1650.) 40) ) aor Specimens with some pravious damage 10 Number of esis 2 3 Aciualipredicted capacity 6sere eee MONOTONIC LOADING IN CONCRETE Fig. 5.41. Comparison of ‘actual failure loads for deformed reinforcing bars loaded in shear towards a ree edge, and values ‘predicted from equation (65.36) (after Paschen & Schonhoff, 1983): n = 38, x= 1.06, 9 = 14% where @ is the strength reduction factor (= 0.85), fis the specified cylinder compressive strength of concrete (Ibf/in*), Ae is the projected area (in?) of the 45° lateral failure cone(s) radiating towards the free edge surface from the bearing surface of the anchor shaft(s), reduced for overlapping areas (i.c. ‘multiple anchors) and intersections of the cone(s) with concrete surfaces (see section 5.2.3.1 for a similar determination of the failure cone for tension loading) and A, = 7c (c + d,)/2 for a single anchor of shaft diameter d, located a distance c from the free edge. When ¢ is taken as 1.0, equation (5.34) is intended to predict the average failure load Where spacing and edge distances are not sufficient to allow the full development of the assumed failure surface, a reduction in capacity directly proportional to the reduction in surface area is assumed. The ACI approach assumes a failure angle of 45°, and the maximum capacity is achieved for a member thickness A > c, Anchors in a row parallel to the edge are assumed to have no influence on each other if the spacing is sufficient to prevent overlap of the failure cones (s., = 2c). This assumption also leads to the conclusion that a single anchor is not affected by a comer if the edge distance c perpendicular to the load direction is at least equal to the edge distance ¢ in the load direction. ‘According to Klingner & Mendonca (1982b), equation (5.34) is valid only for ‘fully embedded’ anchors, full embedment being defined as the embedment that will yield a steel failure for tension loading. Klingner & Mendonca (1982b) ‘compared actual and predicted strengths for tests with headed anchors loaded in shear, using equation (5.34) (see Fig. 9.40). 5.3.3.2, Method according to Shaik & Whayong (1985). Based on analyses ‘of test results from MeMackin et al. (1973) and Klingner & Mendonca (1982b) swith headed studs, Shaik & Whayong (1985) propose the empirical equation Vom = 125 FDS cl — db) (6.35) (am = 5:25 (f% e) ON) In the data considered, the concrete strength was 40505270 Ibffin* (£28MPa—~36MPa) and the anchor diameter was 0.75—2in (~19 mm—~51 mm). 5.3.3.3, Method according to Paschen & Schonhoff (1983). The following emipitical equations are proposed, based on analyses of results from tests with straight deformed reinforcing bars (d, = 14—40ram) in concrete of strength fe = 25-45 MPa: for ple < 1.73 Vim = (190 + 0.230) (f,)°% sin 0.91 cx/c) CN) (5.362) Va = (212 + 0.26c7)()° sin (0.91 cfc) (N) (5.36b) AZ WIA. eo 10 12 13 1 valved casasFig. 5.42. Test results for various anchor types loaded in shear towards a free edge, normalized to d = I8mm, hy = 80mm and f = 2014Pa by use of the terms from equation (5.38) (afer Eligehausen & Fucks, 1988) HEADED ANCHORS IN UNCRACKED CONCRETE and for cle, 2 1.73 Bam = (190 + 0.23¢4)(f..)°" ™ (8.378) = 212 + 0.264) 9 m (5.37) where ¢, is the distance from anchor to free edge parallel to shear load and ‘ys the distance from anchor to free edge perpendicular to shear load. Fig. 5.41 (Paschen & Schénhoff, 1983) shows a histogram of actual~predicted capacity ratios for equation (5.36). 5.3.3.4. y method according to Bligehausen & Fuchs (1988) and Fuchs (1990). (This method is generally referred to in the source literature as the -« method. It has been renamed the y/ method to conform with CEB notation.) ‘Based on regression analyses of 147 tests with headed, expansion and bonded anchors, Eligehausen & Fuchs (1988) and Fuchs (1990) propose the following ‘equation for the calculation of ayerage ultimate failure load of a single anchor loaded in shear towards the edge Vas = d*F2 (ld )9%Ch (6.380) Vim = LL df25 (Mgeid)?7e!> (.38b) ‘The database contains details of anchors of diameter 8—50 mm and of concrete strength approximately 10-50 MPa. The authors propose that this equation is valid for embedment depths 4d < hr < 8d and member thickness h = 14d. For anchors of embedment length hg ~ 4d, equation (5.38a) simplifies to Vom = 13d p5c'4 (6.392) Vem = 14d pS 6.390) Figure 5.42 (Bligehausen & Fuchs, 1988) compares the failure loads predicted by equation (5.38) and the database used for the. regression analysis. The thickness of all test specimens was greater than 1.4c; The test results were normalized to a concrete strength f, = 20MPa, a diameter d = 18 mm and an embedment depth hy = 80mm by use of factors (20/f,)°5, (18/d)°$ and (has/d)°?. Fig. 5.42 is therefore valid only for judging the accuracy of the term c' relative to the database. ‘The y method (see section 5.2.3.3) has been extended by Zhao, Fuchs & Hligehausen (1989) to shear loading by the introduction of a factor to reduce the failure load for member thicknesses less than 1.4¢ 200 © Non-headed anchor Equation (5.98) fa 9 Headed stud s 2 100 ~ 3 ° 2 | I . tj i i | "fe mH 708 300 300 a aMONOTONIC LOADING IN CONCRETE Fig. 5.43. Test results for 1 expansion anchors loaded in Shear tonards a fee edge 1s0/- v nm and exhibiting lazeral rolage concrete-cone failure, of | = plotted for various member 73 thicknesses as a function of gal (og ‘edge distance (after Zhao et al, 1989) dae \e “+ a8 “| ae 700 200 300 00 Fig. 5.44, Comparison of ‘equation (5.41) with test rresulis for expansion anchors loaded in shear fa 1 towards a free edge lo 8 | ‘exhibiting lateral concrete- a Pew] ‘cone failure: rest results have been normalized by the °° a] o 3 value Vig predicted by Equation (541) ‘equation (5.38) and are plowed as a function of hc ~ a4 (after Zhao et al., 1989). ] Via is the average failure lead for a large member dept h = 1-40) o| L 0 rs 35 32 is 20 ne ValVonn Fig. 5.45. Comparison of equation (5.42) with test hhave been normalized by the results for expansion anchor 2g (groups loaded in shear towards a free edge exhibiting lateral concrete- 8 [ uaton 42) ew vale Vay predicted by 1 equation (5.38), and are Vageore!*Mam ‘cone failure: test results v plotted as a function of sic ~~ = (after Eligehausen & Fuchs, - fe 1988) ° t 2 a Fig. 5.46. Anchors in a row parallel to an edge loaded eccentrically (see equation (5.44)) (after Bligehausen & Fucks, 1988) 6Fig. 5.47, Failure mode of «a single anchor loaded in shear when located in. (a) a corner; (b) a narrow member HEADED ANCHORS IN .UNCRACKED CONCRETE Vasa = VV (5.40) Ve TL 4e)°" = 1.0 6.41) where V,. is the average failure load for a single anchor loaded in shear, from equation (5.38). ‘The data used to develop this relationship come from tests with torque- controlled expansion anchors. Zhao et al. (1989) assume that the results are equally valid for headed anchors. Fig. 5.43 plots test results against curves for various values of h resulting from equation (5.40); Fig. 5.44 shows the same test results normalized by the average failure load with h = 1.4c from equation (5.38) arid plotted as a function of h/c, The curve given by equation (6.41) is plotted for comparison. Eligehausen & Fuchs (1988) have proposed a further factor ¥, to account for anchor spacing. Similarly to the approach proposed for headed anchors in tension, a critical value s,,*is assumed to develop the full strength of anchors exhibiting lateral concrete-cone failure. This leads to the following relation for two anchors loaded in shear towards a free edge Vagroup = WsKoon (5.42) Ve 1+ sls, $2 (5.43) where s = 3¢ and Vig is the average failure load for a single anchor Toaded in shear, from equation (5.38). Figure 5.45 (Bligehausen & Fuchs, 1988) gives test data for groups of two anchors loaded toward a free edge normalized with the predicted load for a single anchor from equation (5.38) and plotted as @ function of s/c. The strength predicted by equation (5.42) is plotted for cemparison. This approach is extended to a larger number of anchors with an eccentrically applied load (see Fig. 5.46) by substituting the spacing of the outermost anchors for and introducing the term y, (see séction 5.2.3.3) as follows:(Eligehausen & Fuchs, 1988; Fuchs, 1990) Vagioap = Vee Varn (6.44) % = ssa sn 6.45) Ye = UG + Deh) : 6.46) where sq = 3c, 5; is the spacing of cutermost anchors’ (individual anchor spacing § < 3c), ¢ is the eccentricity of load with respect to the centre of gravity of the anchor group (e < s,/2), n is the number of anchors ina row parallel to the loaded concrete edge, and V,,q is the average failuite load for a single anchor loaded in shear, from equation (5.38). No experimental data for larger anchor groups with/without eccentric loading are available to verify the accuracy of equation (5.44). For the case of anchors located near a corner and loaded in shear towards one edge (see Fig. 5.47(a)), Bligehausen & Fuchs (1988) predict a reduction in the failure load given by equation (5.38) if the distance to the adjacent edge is less than 1.5c. The boundary case of an edge distance c, = 0 is ee @ ©MONOTONIC LOADING IN CONCRETE : 3 7 ' i se Swati 2 7 8 a Siete i =} et etcas ies : f typ dos 8 Hs yk, aap ol 7 ol i ° 05 1 18 ° Os 76 1 20 25 calor cxier Fig. 5.48 (above left). Test assumed to result in a reduction of the failure load by 70%. This leads to resuls for expansion anchors located in a corner Vessornee = WeVam 6.47) ‘and loaded in shear (Fuchs, = el) & Recah aa cane hans ts 0.3 + O.Feqleq) $1 (6.48) been normalized with the where cq = 1.5¢, and V9 is the average failure load for a single anchor value of ¥.,from equation joaded in shear, from equation (5.38). Limited data from tests with expansion (5.38). (after Fuchs, 1990) anchors located in a corer (Fuchs, 1990) are presented in Fig. 5.48. The experimental failure loads are normalized using equation (5.38) and plotted Hig. 5.49 (above righ) as a function of c;/cj. For the tested case c, = cy, equation (5.47) gives @ gi ‘anmparison of fst results et Cameron Schonnosy Yasser = 0.17 Var f Po ictemmacion (330): A further case to be considered with respect to the ¥ method is that of eee hese on deformed a anchor located in a narrow member, as shown in Fig. 5.4700). The reinforcing bars loaded in expression provided for this case. (Rehm et al,, 1988) is Shear ina narrow member: the ulimate loads have been Vasoan enter = ¥eVase 6.49) normalized with the value of yy teeta 5.50) Van from equation (5.38) where ce, = 1.5¢, and V4 is the average failure load for a single anchor loaded in shear, from equation (5.38), with ¢ = ¢). Figure 5.49 plots the results of tests by Paschen & Schénhoff (1983) on deformed reinforcing bars in narrow members. The data have been normalized by the value for V,,_ resulting from equation (5.38) and plotted as a function of cy/cy. In Fig. 5.49, the available data consider only the case of equal edge distances. In the case of unequal edge distances (C2, # ¢,2) where the Gifference is relatively small, Eligehausen & Fuchs (1988) propose that the average values of c,,, and c,2 inserted in equation (5.49) will yield a conservative result for the ultimate load. Fuchs'(1990) cites a lack of test results for anchor groups located in comers Fig. 5.50. Failure pattern a for an anchor group loaded fata in shear towards a free mea? edge (from Rekm et al., 1988) he Fracture crack 2 Row 2—= 0-——-0-===0 Fracture erack 1 Row ie ee 0HEADED ANCHORS IN UNCRACKED CONCRETE, Yaron Load Yayoup: kN f= 25 MPa Hote clearance =2 mm 2 Fig. 5.51. Typical load— displacement curve for an anchorage comprising two anchors aligned perpendicular to a free edge ‘and loaded in shear (after Fuchs & Eligehausen, 1988) Fig. 5.52. Predictions for failure load of an anchor loaded in shear exhibiting lateral concrete-cone failure: f, = 25 MPa, d= Tamm, hie > LA 2 4 Displacement: mm and narrow members, but proposes that the multiplicative effect of the ¥ method will also yield conservative results for this case. As discussed above, a concentric shear load applied to anchors such as headed studs welded to an anchor plate will be distributed approximately equally to all anchors. If the anchor group is loaded in shear towards a free edge, the anchors nearest the edge will fail first (soe Fig. 5.50). The corresponding load carried by the anchorage at this point is given by Rehm et al. (1988) as Va.eroop = Mf + 5f(3.00%,1)1 Vo, (5.51) where ny is the number of anchors arranged one behind another in the direction of the load V. Following failure of the anchors closest to the edge, the shear force is cartied by the more remote anchors, and the maximum load is reached when a parallel lateral failure cone is formed (crack 2 in Fig. 5.50). Eligehausen & Fuchs (988) report that, given typical spacings (¢1,1,¢i,2), ctack 1 has only a slight influence on the rupture load of the group. This implies that the concrete failure load for the group depends only on the row of anchors further from the edge, and can be calculated by the use of equation (5.42), taking ¢ = C1, 28 the distance from the edge. In the case of anchorages in which the anchors are not welded, but project through holes in the anchor plate, some construction clearance is typically provided between the anchors and the plate. In this instance the load—slip behaviour of the anchorage is reported by Eligehausen & Fuchs (1988) to depend significantly on which anchors are initially in contact with the anchor plate, For the simplest case of two anchors aligned perpendicular to the edge with only the anchor nearer the edge initially in contact with the loading plate, the observed load—displacement behaviour of the group is shown in Fig. 4007 a — actato.as (1085) a= Shates wnayong (1988) a= Pasehon a Senannot (1983) o= erganausen Fucts (1888) i 4 3 -200- le 5 400- 180 ‘200 sto Distance to fre edge & mm “ato nMONOTONIC LOADING IN CONCRETE Fig. 5.53. Predictions for 19) - ultimate strength of an ee ‘anchor located ina comer BA exhibiting lateral concrete L { Gone failure C4 | Y 7 Y er | g a 2 E os 4 , : 7 2 7 / ; {AC 389-05 (1995) / —=— Eligehausen & Fuchs (1988) TS Pasenen & Schon (963) °% 1 2 fer Fig. 5.54, Predictions for timate strength of @ group of two anchors exibiing lateral concrete-cone failure = t = Act 6#0-85 1808) = === Eigenauson & Fuchs (198) 1 °9 1 2 3 wc Fig. 5.55. influence of a 10 reemant on the load— displacement Behaviour of headed anchors loaded in shear (from Rehm et al., 1988): fe, ~ 281APa, dy = Tamm, d = 22mm 1 ® S od ote Fa ® % 5 10 1s Displacement: mm nHEADED ANCHORS IN UNCRACKED CONCRETE 5.51 (Bligehausen & Fuchs, 1988). Typically, failure of the conerete in front Of the anchor closer to the edge is followed by load transfer to the second anchor. Since the load is again carried by the anchors in series, the maximum load as given by the second peak in Fig. 5.51 is determined by the anchor further from the edge. Since the anchor further from the edge determines the maximum load, this is theoretically independent of the order in which the anchors come in contact with the plate during loading, and equation (5.42) is still applicable. 5.3.3.5. Comparison of different approaches. Figure 5.52 compares the average ultimate shear loads for single-headed anchors exhibiting lateral “concrete-cone failure predicted by the methods described in sections '5.3.3.1—5.3.3.4, For the usual concrete strength (C25) assumed for this ‘comparison, the equations proposed by Shaik & Whayong (1985) and Fligehausen & Fuchs (1988) are almost identical. According to ACI 349-85 (1985) and Paschen & Schémhoff (1983), the shear capacity of an anchor exhibiting lateral cone concrete failure increases proportionally to c*. This leads to higher failure Joad values than the predictions of other methods for edge distances = 100mm. Except for that of Paschen & Schénhoff (1983), all the methods assume that the failure load is directly proportional to the square root of the ‘compression strength of the conerete. Paschen & Schénhoff (1983) assume proportionality of the failure load to the term f2°. The tests on which they based their conclusions were conducted in concrete with litte strength variation, ‘The predictions for failure load given by the various methods for the case of a single anchor in a corner are compared in Fig. 5.53. The calculated values are divided by the failure load assuming no edge influence, and are plotted as a function of c2/c,. The methods of Paschen & Schénhoff (1983) and Bligehausen & Fuchs (1988) agree quite well for the range (c2/c\) > 0.5. By virtue of the model used to calculate the reduced load, ACI 349-85 (1985) assumes a much smaller influence ‘of edges on the failure load than the other methods. Figure 5.54 compares the predictions of the various methods for anchor groups (two anchors) exhibiting lateral concrete-cone failure. The reduced failure loads are divided by the failure loads for two single anchors without spacing influence as predicted by the proposals, and are plotted as a function ‘of s/c. As with the case of an anchor located in a corner, the ACI 349-85 (1985) equation predicts higher capacity than the relation given by Eligehauscn & Fuchs (1988) over the range (s/c) <3. 5.3.3.6. Reinforcement, The sheat load-bearing capacity of headed anchors located near the edge of a member can be increased by appropriate reinforcement, This is shown in Fig. 3.55 (Rehm et al., 1988), which plots typical load—displacement citrves for headed anchors (shank diameter d = 22mm) with a distance from the edge of ~75mm. While the increased capacity achieved by the reinforced anchorage represented by curve (2) is significant, it is far less than would be expected if complete load transfer to the stirrups on either side of the anchorage had been achieved. The investigators concluded that this is explained by the poor anchorage of the stirrups in the area of the lateral failure cone. Curve (3) demonstrates the significant load increase given by the introduction of hairpin reinforcement around the anchorage. In this case, however, the ultimate load attained was still below the calculated capacity of the hairpin. The investigators teported that subsequent examination of the concrete between the anchor and the hairpin revealed that it had been crushed. Paschen & Schénhoff (1983) conducted similar tests in which the hairpin was placed in direct contact with the anchorage. It was found that the yield strength of the hairpin could be developed given adequate hairpin enchorage aMONOTONIC LOADING IN CONCRETE Fig. 5.56. Recommended reinforcement for anchors near an edge loaded in shear (from Paschen & Schonhoff, 1983): ty is the development lengtit ‘according to CEB-FIP Model Code (1990) ™ hs teem em v vot a ¢ | nad 50 mn | beyond the zone of failed concrete (see Fig, 5.56). Klingner et al. (1982) also conducted tests on anchors in direct contact with hairpin reinforcement They found the proximity of the hairpin to the top surface of the concrete to be a significant factor in the performance of the anchorage. Anchorages with the hairpin placed within 20 mm of the surface showed a far greater increase in both stiffness and strength than anchorages with the hairpin set 50 mm below the concrete surface. The investigators concluded that this is a result of the additional bending moment experienced by the anchor bolt in the latter cas. ‘The initial rigidity of the anchorage is typically not influenced by reinforcement, However, the maximum ioad increases depending on the size and pattern of the ‘reinforcement as discussed, 5.3.3.7. Shear load parallel to the edge. If an anchor close to an edge is loaded with a shear force acting parallel to the free edge, research by Stichting Bouwresearch (1971) indicates that the failure load is about twice as large as can be expected for loading perpendicular to the edge. However, little is known about this phenomenon. 5.3.4, Failure loads — concrete crushing combined with anchor pull-out Headed anchors loaded in shear typically exhibit crushing of the concrete in front of the anchor. For anchors with a relatively small embedment depth, this may lead to a concrete fracture behind the anchor which in turn causes a pull-out failure (see Fig. 5.37(g)). While no theoretical formulations are available at present to predict the failure load associated with this type of failure, it is likely to depend principally on the concrete strength, the embedment depth, the head diameter and the spacing of the anchors, as applicable. Generally, the required embedment to preclude this type of failure is four to five times the diameter of the anchor. For high-strength steels or low-strength concrete and anchor groups, however, this embedment length may net be adequate 5.3.5. Failure loads — concrete spall ‘The primary factors influencing the shear load at which spalling of the concrete occurs are tensile strength of the concrete, flexural stiffness of the anchor shaft, anchor shaft diameter, embedment depth and deformability (E modulus) of the concrete, The corresponding shear capacity is given by Olgeard, Stutter & Fisher (1971), AISC (1978), Klingner & Mendonca (1982b) and Roik (1982) as Vim = O-5AGE) (bf, N) (5.52) Hawkins (1987) proposes the following equation based on statistical analysis of test results Vem = 18.2 (dy? f205 (1S + 1.1 Aye + dw) (Ibe) (5.53) Com 313-1 FoF BBL + LL he + dw) ON)5.4, Combined tension and shear loading Fig. 5.57. Interaction diagram for headed anchors (est results from Bode & Hanenkamp, 1985) HEADED ANCHORS IN UNCRACKED CONCRETE where dw = washer diameter < hy. The tests were conducted on headed studs with-a washer beneath the nut. If no washer is present, dw is taken as zero in equation (5.53) ‘From tests with headed studs, the depth of the spall was found to be typically about 0.4—0.6 times the anchor diameter (Fuchs, 1990). Based on theoretical considerations, the following equation was proposed for the spalling load Vom = As(O-Uf, + 2.%fce) — (N) 6.54) Fuchs (1990) limits the applicability of equation (5.54) to anchors with an embedment depth of at least five times the anchor shaft diameter. For embedment depths of between three and five times the shaft diameter, a deduction factor is proposed Uy = O2hegld <1 (5.55) Investigations by Eligehausen & Fuchs (1988) have indicated that the ultimate oad attained by the anchor in the case of steel failure is not greatly affected by the nature of the concrete spall. 5.4.1, Steel failure "The steel failure mechanism for headed anchors in combined tension and shear is characterized by yielding and fracture of the anchor shaft due to combined shear, tension and flexural stresses. 5.4.1.1. Meihod according to ACI 349-85 (1985) (shear— friction theory). ‘The shear strength of single anchors or anchor groups with full embedment is calculated by use of shear—friction theory. This approach is valid for steel failure only. Equation (5.56) is derived frome the design equation in ACT 349-85, (1985), with a strength reduction factor @ = 1.0. N+ Vin = Ady (5.56) ihre bis ia le: pr eae | mae] ||, ‘i Lrwlo]|ofe = 22 | 100 a . ! Pet Hh 2 |io| 2 | "4 a TSS 3 N oe 2 cel BS \ Equation (5.61) Pery nee p a 0.50] A laa a \ oO 0.25) - quan (680) °o 0.25, 0.50 0.75 1.00 = 1.50 VoseeV Ves! 6MONOTONIC LOADING IN CONCRETE 6 where 1 is the coefficient of friction (~ 0.55~—0.9, depending on the location of the anchor plate in relation to the concrete surface) (see section 5.3.2.1), A, is the stress area for all threaded anchors A, = (14) (doug = 0.9743/n)? 6.57) ris the number of threads per inch (254mm), From equation (5.56), the ACT approach to interaction gives a straight-line interaction diagram 5.4.1.2. Methods not based on shear—friction theory. The strength of steel bolts and studs subjected to tension and shear loads is commonly represented as an elliptical interaction curve. The interaction equation that defines the ultimate load thus takes the form WIN + VY" S10 6.58) cctypically varies between 5/3 (McMackin et al., 1973) and 2 (Thurner, 1985; Shaik & Whayong, 1985). 5.4.2. Concrete failure To calculate the failure load under combined tension and shear loading, three approaches are found in the literature, These may be characterized as straght- line, trilinear and elliptical functions. 5.4.2.1, Straight-line function. Based on test results for wedge-type expansion anchors, Johnson & Lew (1990) propose a straight line as a lower bound for the interaction diagram (see Fig. 5.57) (NIN,) + (WV) = 1.0 6.59) 5.4.2.2, Trilinear function. Bode & Roik (1987) propose a trilinear function for headed anchors NIN, <1 (6.604) Wy st (6.600) (WIN,) + (VIV,) 12 (5.60¢) where N and are the applied tension and shear loads respectively and Ny and V, are the ultimate tension and shear loads respectively. Figure 5.57 shows the results of tests on headed anchors in which the angle of the load was systematically varied between a = 0° (pure shear) and « = 90° (tension). The test results are presented in the form of an interaction. diagram. The measured failure loads are related to the 5% fractile of the theoretical values as calculated by use of the y method (see sections 5.2.3.3 and 5.3.3.4). ‘Some tentative experimental investigations using quadruple fastenings indicate that the interaction equations (equation (5.60)) can also be used for anchor groups (Rehm et al., 1988). However, the authors concluded that further rescarch is needed to clarify the bearing mechanism and to establish the influence of such major parameters as number and diameter of fastening elements, anchorage depth and anchor spacing on the load—-displacement behaviour and the ultimate load. In particular, they proposed tests designed to establish the transition in failure mode for anchor groups loaded in varying degrees of tension and shear. No experimental or theoretical work has yet been conducted in this area 5.4.2.3. Elliptical functions, MeMackin et al. (1973), Shaik & Whayong (1985) and Cook & Klingner (1989) propose an elliptical function to describe the strength of anchors loaded in tension and shear WINS + IV," = 1.0 6.61) where a is determined from analysis of test results. Shaik & Whayong (1985) propose a = 2, which gives a circular functionHEADED ANCHORS IN-UNCRACKED CONCRETE 120) — 0} “Tonsion: KN Fig, 5.58. Tenson—shear | interaction for castin-place anchors (Cook & Klingner, °0 40 60 1989) Shear: KN (Fig. 5.57); MceMackin et al. (1973) and Cook & Klingner (1989) propose @ = 5/3, which leads to an interaction diagram in the form of a flattened ellipse (Fig. 5.57). PCI (1985) proposes the use of a = 4/3 for anchors used. ‘ in precast construction. 2 ‘Cook & Klinger (1989) measured anchor tensile forces directly for multiple- ‘ anchor connections. In the calculation of the anchor shear forces, the coefficient of friction was known and the equilibrium relations discussed in section 5.6.4 were used. As shown in Fig. 5.58, an elliptical tension—shear interaction (see also equation (5.61)) relation was fouind to provide a reasonable and generally conservative fit to the tést data. A linear tension—shear interaction relation was found to be conservative. 5.5. Bending of the To determine the ultimate strength of the steel-to-concrete connection shown baseplate in Fig. 5.59, three separate capacities must be considered (@ the flexural capacity of the attached member ' ( the flexural capacity of the steel baseplate (0) the capacity of the fastener. Fig. 5.59. Typical steel-to- a ttomont concrete connection oe Attaches membor Fastener ( nMONOTONIC LOADING IN CONCRETE 5.6. Behaviour of multiple-anchor ductile attachments to concrete If the connection is proportioned so that yielding occurs in the attached ‘member, the baseplate or the anchors before failure of the concrete, then the connection can be described as ductile. ‘The flexural capacity of the attached member can be calculated by the same means as used for structural steel design, and is not discussed here. The flexural capacity of the baseplate and the capacity of the fastener itself are discussed in this section and section 5.6 respectively. In the case of a perfectly rigid baseplate, the compression resultant is located. at the toe of the baseplate. In the case of a flexible baseplate, tests by Cook & Klingner (1989) show that the location of the compressive reaction moves closer to the tensile anchors, and is given by the condition of moment equilibrium between the applied moment and the tensile resultant of the anchors, ‘As shown in Fig. 5.60, the overhanging portion of the plate on which the compressive reaction acts is essentially a cantilever fixed at its intersection with the Compressive element of the attached member, and loaded by a movable concentrated load (the compressive reaction). The boundary conditions for this cantilever require the fixed support to rotate and displace. ‘The free end can rotate but cannot dispiace if the compressive reaction is at the toe of the plate, and can neither rotate nor displace if the compressive reaction moves in from the toe of the baseplate. ‘As shown in Fig. 5.60, the behaviour of a flexible baseplate can be described as follows. (@ Initially, the baseplate rotates as a rigid body, pivoting about the toe of the plate. () As the compressive reaction increases, the portion of the baseplate adjacent (o the compressive element of the attached member reaches the yield moment M, of the baseplate. This causes the compressive reaction C to move inwards towards the compression element. The smallest distance qi, between the compressive reaction and the ‘compression element of the attached member can be determined from Sein = MyIC 6.62) (6) With a further increase in the compressive reaction, the baseplate forms a plastic hinge M, and the compressive reaction C moves away from the compression element. The longest distance that the compressive reaction moves away from the support is determined from X= MIC (5.63) To locate the compressive reaction conservatively, it can be considered to be at a distance xj, determined by equation (5.62) from the edge of the compression element of the attached member. ‘The most extensive currently available investigation of the behaviour of ductile ‘multiple-anchor attachments to concrete is that of Cook & Klingner (1989), whose principal findings are discussed in this section. The experimental programme included the following types of test (a) coefficient of friction (©) ultimate load: two-anchor rigid baseplate, four-anchor rigid baseplate, six-anchor rigid baseplate and six-anchor flexible baseplate. 5.6.1. Coefficient of friction In the tests of Cook & Klingner (1989), the average mean value for the coefficient of friction between a surface-mounted stee! plate and hardened concrete was 0.43, with a standard deviation of 0.09. The coefficient of friction was not significantly affected by the surface condition of the concrete, theFig. 5.60. Effect of baseplate flexibility on the location of the compressive reaction (Cook & Klingner, 1983): (a) compressive reaction at 10¢; (b) compressive reaction shifted in from toe Fig. 5.61. Typical load— displacement diagrams for ssit-anchor rigid baseplate test dominated by anchor shear (Cook & Klingner, 1989) HEADED ANCHORS IN UNCRACKED CONCRETE Meaty — of BREE MaMy je ce HL “magnitude of the compressive force or ‘digging in’ of the toc of the rigid baseplate to the concrete. For computation of capacity, the authors recommend that the coefficient of friction « be taken as 0.40, with a strength reduction factor ¢ of 0.65. . ) ® 5.6.2. Uliimate capacity of multiple-anchor attachments to concrete ‘All ultimate-load specimens of Cook & Klingner (1989) failed by yielding and fracture of the anchors. Three typical load—displacement curves (Fig. 45.61) show the applied shear as a function of the total displacement (the square root of the sum of the squares of the horizontal slip and the vertical displacement at the location of the outer row of tension anchors). Even for connections dominated by anchor shear, the anchors underwent significant inelastic deformation before failure. 5.6.3. Combined tension and shear interaction relationships ‘The descriptions given in section 5.4 are applicable. More details of multiple- anchor connections under combined tension and shear load are given in section 5.4.2.3, 5.6.4, Distribution of tension and shear among anchors Results of the multiple-anchor attachment tests of Cook & Klinger (1989) indicated that (@ tension and shear forces in the anchors redistribute inelastically as required to maintain equilibrium with the applied loading (b) for connections dominated by moment (high eccentricity of the applied load) the anchors away from the toe of the baseplate attain their full tensile strength (©) for connections dominated by shear (low eccentricity of the applied load), the ultimate strength of the connection is not sensitive to the distribution of tension in the anchors es + 4 : | es Displacement mmBinitaapuaete te MONOTONIC LOADING IN CONCRETE Table 5.2. Ranges of behaviour for ductile ‘multiple-anchor connections (Cook & Klingner, 1989): T, = the tensile strengih of the anchor @) the initial distribution of anchor tension (before inelastic redistribution) does not affect the ultimate strength of the connection. 5.6.5. Analytical basis for prediction of the strength of ductile mutiple-anchor connections According to Cook & Klingner (1989), the behaviour of a ductile multiple- anchor connection can be separated into three distinct ranges (Table 5.2). (@ Range I. if the shear strength provided by the frictional force (developed from the compressive reaction produced by the applied moment) is larger than the applied shear, anchors are not required for shear. The anchors in the tension zone can be assumed to develop their full tensile strength for moment resistance, (®) Range 2. If the shear strength provided by the frictional force and by the anchors in the compression zone exceeds the applied shear, the anchors in the tension zone can be assumed to develop their ull tensile strength for moment resistance. (© Range 3. If the shear strength provided by the frictional force and by the anchors in the compression zone is less than the applied shear, the anchors in the tension zone must transfer the remaining shear load. The strength of the anchors in the tension zone is limited by their tension—shear interaction. The transitions between these three ranges of behaviour coincide with two critical values of shear load eccentricity e, equal to the moment—shear ratio M/Y of the applied loading at the surface of the concrete, The first critical eccentricity e” represents the transition between range 1 and range 2 behaviour, and corresponds to the point at which the applied shear load is equal to the frictional force. For eccentricities larger than e’, the connection does not slip and no shear anchors are required. For eccentricities smaller than e’, the connection slips and shear anchors must be provided. ‘The second critical eccentricity e” represents the transition between range 2 and range 3 behaviour, and corresponds to the point at which the applied shear load is equal to the sum of the frictional force and the shear strength of the anchors in the compression zone. For eccentricities larger than e”, ‘the anchors in the tension zone can be assumed to develop their full tensile strength for moment resistance. For eccentricities smaller than e”, the anchors in the tension zone carry both tension and shear. If no.anchors are provided in the compression zone, range 2 behaviour is not applicable, and e" =e". Range 3 Range 2 Range 1 Connection stips Connection slips Connection does not slip Anchors in the compression zone | Anchors in the compression zone | Anchors are not requied for shear are at their maximum shear strength 77, transfer shear and can achieve their maximum shear strength ¥7, Anchors in the tension zone are in | Anchors in the tension zone can | Anchors in the tension zone can combined tension and shear and | achieve their maximum tensile achieve their maximum tensile ccan achieve their maximum strength in combined tension and shear strength T, strength T, ‘Shear load eccentricity ¢ or moment—shear ratio M/V ° 80 e e °HEADED ANCHORS IN-UNCRACKED CONCRETE 5.6.6, Equations to predict the capacity of ductile imultiple-anchor connections The strength of connections dominated by shear depends on the tension— shear interaction of the anchors. As-derived equations are given below for elliptical and linear interaction (Cook & Klingner, 1989) For elliptical tension—shear interaction, the anchor shear strength is given by Vy = (T3 — TH)? (5.64) For the more conservative linear tension —shear interaction, the anchor shear strength is given by Va = (To — Tr) (5.65) where V, is the shear strength of an anchor in combined tension and shear and + is the ratio of the shear strength of the anchor to the tensile strength of the anchor (0.5 for cast-in-place and adhesive anchors and 0.6 for undercut anchors, see section 5.3.2.2) ‘The critical eccentricities e” and e” may be determined by the conditions of equilibrium, The following formulations for ¢’ and e” ate applicable to connections with multiple rows of anchors if d is taken as the distance from the compressive reaction C to the centroid of the anchors in the tension zone, ris taken as the number of rows of anchors in the tension zone, m is taken ‘as the number of rows of anchors in the compression zone and T, is taken as the tensile strength of a row of anchors. ‘At the minimum eccentricity e”, the applied shear load V is equal to the frictional force #C ef = dip (5.66) where e’ is the minimum eccentricity for multiple-anchor connections without shear in the anchors, iis the coefficient of friction between steel and concrete, and d is the distance from the compressive reaction to the centroid of the anchors in the tension zone. ‘At the minimum eccentricity ¢”, the applied shear load V is equal to the sum of the frictional force C and the shear strength of the rows of anchors in the compression zone myT e” = ndl(np + my) 6.6 where e” is the minimum eccentricity for multiple-anchor connections without combined tension and shear in the anchors, n is the number of rows of anchors in the tension zone, m is the number of rows of anchors in the compression zone, + i the ratio of the shear strength of the anchor to the tensile strength of the anchor, and . and d are as defined for equation (5.66). ¢” reduces to e’ when no anchors are provided in the compression zone. 5.6.6.1. Distribution of tension. Fot connections with more than one row of anchors in the tension zone, the distribution of tension cannot be adequately predicted by traditional analysis methods. Provided that sufficient anchors are present to satisfy cquilibrium, the connection will perform satisfactorily However, the issue of available inelastic deformation capacity must be addressed. The above analysis is based on the assumption that materials have infinite plastic deformation capacity after yield: this is not in fact the case. Cook & Klingner (1989) have proposed that the distance d; between the inner row of anchors and the compression reaction should not be less than ~10% of the distance d, (4, > 0.10 d, in Fig. 5.62). This is an attempt to ensure that the tensile strain in the inner row of anchors ¢ will be at least 0.01 when the tensile strain in the outer row of anchors ¢ reaches its maximum value ¢,. Anchor materials typically have a specified minimum elongation requirement of at least 10% in 50mm. This represents an ultimate a